Most research on constipation has focused on dietary fiber intake. Here, we examined the intake of water and magnesium, nutrients possibly associated with constipation, as well as that of dietary fiber in relation to constipation.
A total of 3835 female Japanese dietetic students aged 18–20 years from 53 institutions in Japan.
Dietary intake was estimated with a validated, self-administered diet history questionnaire. Functional constipation was defined using the Rome I criteria.
The prevalence of functional constipation was 26.2%. Neither dietary fiber intake (mean=6.4 g/4186 kJ) nor intakes of total water and water from fluids were associated with constipation. Conversely, low intake of water from foods was associated with an increasing prevalence of constipation. In comparison with women in the first (lowest) quintile, the multivariate adjusted odds ratio (OR) (95% confidence interval (CI)) for women in the second, third, fourth, and fifth quintiles were 0.72 (0.57, 0.90), 0.78 (0.62, 0.98), 0.71 (0.56, 0.89), and 0.77 (0.61, 0.97), respectively (P for trend=0.04). Additionally, low magnesium intake was associated with increasing prevalence of constipation. Compared with women in the first quintile, the multivariate adjusted OR (95% CI) for women in the second, third, fourth and fifth quintiles were 0.70 (0.56, 0.88), 0.75 (0.60, 0.95), 0.73 (0.58, 0.92) and 0.79 (0.63, 0.996), respectively (P for trend=0.09).
Low intakes of water from foods and magnesium are independently associated with an increasing prevalence of functional constipation among a population whose dietary fiber intake is relatively low.
Constipation is a common health problem (Wong et al., 1999; Pare et al., 2001; Garrigues et al., 2004; Higgins and Johanson, 2004), and diet is considered a major modifiable lifestyle factor associated with this condition (Locke et al., 2000; Talley, 2004). The favorable effect of dietary fiber on constipation is widely accepted and several (Dukas et al., 2003; Sanjoaquin et al., 2004), although not all (Campbell et al., 1993; Towers et al., 1994; Murakami et al., 2006), observational studies have indicated an inverse relation between dietary fiber intake and constipation. However, while most previous studies have defined constipation according to the infrequency of bowel movement (Campbell et al., 1993; Towers et al., 1994; Dukas et al., 2003; Sanjoaquin et al., 2004) or the subjective perception of patients (Murakami et al., 2006), a consensus definition of constipation consists of straining, hard stools and incomplete evacuation in addition to infrequency (Rome criteria) (Whitehead et al., 1991).
Other nutrients that might be associated with constipation include water and magnesium. Low intake of water can reduce the water content of stools and hence lead to constipation (Arnaud, 2003), although the potential benefit of an increase in intake is unknown (Klauser et al., 1990; Anti et al., 1998; Young et al., 1998; Chung et al., 1999). Magnesium might form sulfate or citrate salts that would promote fluid retention in the digestive tract and indirectly alter motility, and thereby act as a light laxative (Saez, 1991). To our knowledge, however, no observational studies have investigated the intake of water and magnesium in relation to constipation.
The aim of this cross-sectional study of young Japanese women was to examine associations between dietary fiber, water and magnesium intake, as assessed with a previously validated, self-administered diet history questionnaire (DHQ) (Sasaki et al., 1998a, 1998b, 2000), and functional constipation as defined according to the Rome criteria (Whitehead et al., 1991).
Subjects and methods
Subjects and survey procedure
The study was based on a self-administered questionnaire survey of a wide range of dietary and non-dietary behaviors among dietetic students (n=4679) from 54 universities, colleges and technical schools in 33 of 47 prefectures in Japan. Staff at each institution distributed two questionnaires on dietary habits (DHQ) and other lifestyle items during the previous month to students during an orientation session or a first lecture designed for freshman students entering dietetic courses in April 2005; in most institutions, this was carried out within 2 weeks after the course began. Students filled out the questionnaires during the session, lecture, or at home and then submitted the completed forms to staff at each institution as soon as possible. A third questionnaire on lifestyle during the previous 6 years (i.e. junior high school and high school) was also distributed and answered in similar fashion; in most institutions, this was carried out within 4 weeks after the course began.
The staff at each institution checked the responses as soon as possible according to the survey protocol. When missing answers or logical errors were identified, the student was asked to complete the questionnaire again. The staff at each institution mailed the questionnaires to the survey center. Staff at the survey center checked the answers again and, when necessary, returned problematic questionnaires to staff at the respective institution, and the student was asked to complete the questionnaires again. All questionnaires were thus checked at least once by staff at the respective institution and by staff at the survey center. Most surveys were completed by May 2005. The protocol of the study was approved by the Ethics Committee of the National Institute of Health and Nutrition.
A total of 4286 students (4066 women and 220 men) answered all three questionnaires (response rate=91.6%). For the purposes of the current analysis, we selected female subjects aged 18–20 years (n=3967). We then excluded from these 3967 women those who were in an institution where the survey had been conducted at the end of May (n=97), those with extremely low or high energy intake (<2093 or >16744 kJ/day) (n=23), and those with missing information on the variables studied (n=24). As some subjects were in more than one exclusion category, the final analysis sample comprised 3825 women. Further exclusion of subjects with intentional dietary change within the preceding year (n=649), those habitually using oral laxatives (n=231), or both did not materially alter the findings, and these were therefore included in the analyses.
Dietary habits during the previous month were assessed using a previously validated, self-administered DHQ (Sasaki et al., 1998a, 1998b, 2000). This is a 16-page structured questionnaire that consists of the following seven sections: general dietary behavior; major cooking methods; consumption frequency and amount of six alcoholic beverages; consumption frequency and semi-quantitative portion size of 121 selected food and non-alcoholic beverage items; dietary supplements; consumption frequency and semi-quantitative portion size of 19 staple foods (rice, bread and noodles) and miso soup (fermented soybean paste soup); and open-ended items for foods consumed regularly (once/week) but not appearing in the DHQ. The food and beverage items and portion sizes in the DHQ were derived primarily from data in the National Nutrition Survey of Japan and several recipe books for Japanese dishes (Sasaki et al., 1998a).
Estimates of dietary intake for 147 food and beverage items, energy, total, soluble and insoluble dietary fiber, total water, water from fluids, water from foods, and magnesium, were calculated using an ad hoc computer algorithm for the DHQ, which was based on the Standard Tables of Food Composition in Japan (Science and Technology Agency, 2000). Information on dietary supplements and data from the open-ended questionnaire items were not used in the calculation of dietary intake. Dietary fiber was determined by an enzymatic-gravimetric procedure (modified Prosky method) (Science and Technology Agency, 2000) from the intake of 86 fiber-containing foods in the DHQ. Total water was defined as the sum of water from all 147 food and beverage items. Water from fluids was defined as the sum of water from all beverages, milks, juices, and soups and water, whereas water from foods was defined as the sum of water from all other food items. Although we calculated magnesium intake from foods and drinks only, and not from dietary supplements, no subjects used magnesium supplements and only 14 (0.4%) used multimineral supplements, rendering it unlikely that dietary supplementation had a major impact on the findings. Detailed descriptions of the methods used for calculating dietary intake and the validity of the DHQ have been published elsewhere (Sasaki et al., 1998a, 1998b, 2000). The Pearson correlation coefficient between DHQ and 3-day estimated dietary records was 0.48 for energy among 47 women (Sasaki et al., 1998a). In addition, the Pearson correlation coefficients between DHQ and 16-day weighed dietary records were 0.69 for total dietary fiber, 0.62 for soluble dietary fiber, 0.70 for insoluble dietary fiber, 0.25 for total water, 0.25 for water from fluids, 0.64 for water from foods and 0.57 for magnesium in 92 women (S Sasaki, unpublished observations, 2006).
A constipation questionnaire was developed based on a previous study (Garrigues et al., 2004) and incorporated into the 20-page questionnaire for lifestyle during the previous 6 years. We used the definition of functional constipation recommended by an international workshop on the management of constipation (Rome I criteria) (Whitehead et al., 1991). Although the Rome I criteria were modified in 1999 to the Rome II criteria (Thompson et al., 1999), epidemiologic studies have consistently shown that the latter may be too restrictive for the diagnosis of constipation (Pare et al., 2001; Garrigues et al., 2004), and we therefore used the former. The following four questions were used to assess Rome I-defined functional constipation: (1) Do you strain during a bowel movement? (2) Do you feel an incomplete emptying sensation after a bowel movement? (3) How often are your stools hard? and (4) How many bowel movements do you usually have each week? These questions referred to the last 12 months. For questions 1–3, four answers were offered: never, sometimes (<25% of the time), often (25% of the time) and always. Functional constipation was defined as meeting two or more of the four criteria (an answer of often or always to questions 1–3 and <3 bowel movements per week (question 4)).
In the questionnaires, subjects reported body weight and height, residential area, current smoking (yes or no) and oral medication usage (yes or no). Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). We classified BMI into three categories (<18.5, 18.5–24.9 and 25 kg/m2) according to the Japan Society for the Study of Obesity (Matsuzawa et al., 2000). The reported residential areas were grouped into six categories (Hokkaido and Tohoku; Kanto; Hokuriku and Tokai; Kinki; Chugoku and Shikoku; and Kyushu) based on the regional blocks used in the National Nutrition Survey in Japan (Ministry of Health, Labour, and Welfare, 2004) (hereafter referred to as ‘residential block’). The residential areas were also grouped into three categories according to population size (city with a population 1 million; city with a population <1 million; and town and village) (hereafter referred to as ‘size of residential area’).
Additionally, subjects reported the time when they usually went to bed and arose in the morning, which was used to calculate sleeping hours, and the frequency and duration of high- and moderate-intensity activities, walking, and sedentary activities. For subjects whose recorded total hours were <24 h, unrecorded hours were assumed to be spent on sedentary activities. For subjects whose recorded total hours were >24 h, the total number of hours spent daily were proportionately decreased to equal 24. Each activities was assigned a metabolic equivalent (MET) value from a previously published table; 0.9 for sleeping, 1.5 for sedentary activity, 3.3 for walking, 5.0 for moderate-intensity activity and 7.0 for high-intensity activity (Ainsworth et al., 1993, 2000). The number of hours spent per day on each activity was multiplied by the MET value of that activity, and all MET-hour products were summed to give a total MET-hour score for the day. This score essentially corresponds to the number of kilojoules per kilogram of body weight expended by an individual during the day. The standard value of basal metabolic rate for Japanese people is also expressed as the number of kilojoules per kilogram of body weight expended by an individual during the day. Physical activity level was then calculated by dividing total MET-hour score (kJ/kg of body weight/day) by the standard value of basal metabolic rate for Japanese women aged 18–29 years (99 kJ/kg of body weight/day) (Ministry of Health, Labour, and Welfare, 2005).
Associations between functional constipation (the dependent variable) and energy-adjusted (/4186 kJ) intakes of total, soluble and insoluble dietary fiber, total water, water from fluids, and water from foods and magnesium were examined. We calculated both crude and multivariate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for functional constipation for each quintile category of dietary variable using logistic regression analysis. Multivariate adjusted ORs were calculated by adjusting for BMI (three categories), residential block (six categories), size of residential area (three categories), current smoking (two categories), current alcohol drinking (two categories (yes or no) because of extremely low alcohol intake: mean=0.8 g/day), oral medication usage (two categories), physical activity level (quintiles) and energy intake (quintiles). We further conducted multivariate analyses including dietary fiber, water and magnesium simultaneously in order to investigate the independent associations with constipation. As results for the crude and multivariate analyses were similar for all variables analyzed, we present here only those derived from the multivariate models. Trend of association was assessed by a logistic regression model assigning scores to the levels of the independent variable. All statistical analyses were performed using SAS statistical software, version 8.2 (SAS Institute Inc., Cary, NC, USA). All reported P-values are two-tailed, and a P-value of <0.05 was considered statistically significant.
Basic characteristics of the subjects are shown in Table 1. Mean total dietary fiber intake was 6.4 g/4186 kJ, mean total water intake was 1025 g/4186 kJ and mean magnesium intake was 119 mg/4186 kJ. A total of 1002 women (26.2%) were classified as having constipation. Table 2 shows the multivariate adjusted ORs for constipation by quintiles of dietary variables. Dietary fiber intake was not associated with constipation. Further, no association was seen for total water intake or intake of water from fluids. However, low intake of water from foods was associated with increasing prevalence of constipation. In comparison with women in the first (lowest) quintile of intake of water from foods, the multivariate adjusted OR (95% CI) for women in the second, third, fourth and fifth quintiles were 0.72 (0.57, 0.90), 0.78 (0.62, 0.98), 0.71 (0.56, 0.89) and 0.77 (0.61, 0.97), respectively (P for trend=0.04). Low magnesium intake was also associated with increasing prevalence of constipation. Compared with women in the first quintile of magnesium intake, the multivariate adjusted OR (95% CI) for women in the second, third, fourth and fifth quintiles were 0.70 (0.56, 0.88), 0.75 (0.60, 0.95), 0.73 (0.58, 0.92) and 0.79 (0.63, 0.996), respectively (P for trend=0.09). Including dietary fiber, water and magnesium intake simultaneously in the models generally attenuated the association between dietary intake and constipation (see multivariate and nutrient-adjusted ORs in Table 2). However, these analyses did not materially change the relations of intake of water from foods and intake of magnesium to constipation, suggesting that both are independently associated with an increasing prevalence of constipation.
To our knowledge, this study is the first to examine dietary fiber, water and magnesium intake in relation to Rome I-defined functional constipation. After controlling for a series of potential confounding factors, we found that a low intake of water from foods and magnesium was associated with an increasing prevalence of functional constipation. In contrast, no association was seen for dietary fiber, total water and water from fluids.
The prevalence of Rome I-defined functional constipation in the present group was 26.2%. A similar prevalence by these criteria has been reported in Canadian (21.0%) (Pare et al., 2001) and Spanish (28.6%) (Garrigues et al., 2004) women, whereas a somewhat smaller ratio was seen in elderly Singaporean women (10.5%) (Wong et al., 1999).
Increased intake of dietary fiber is widely considered to protect against constipation, and several studies have indeed found an inverse relation between dietary fiber intake and constipation (Dukas et al., 2003; Sanjoaquin et al., 2004). Here, however, in common with several other studies (Campbell et al., 1993; Towers et al., 1994; Murakami et al., 2006), we failed to find such an association. A possible explanation for this is that the dietary fiber intake of most subjects was too low to have a protective effect. Estimated intake in the present study (mean=11.8 g/day) was, however, comparable to that observed in women aged 18–29 years in the Japanese National Nutrition Survey (mean=12.0 g/day) (Ministry of Health, Labour, and Welfare, 2004).
Although we saw no relation between the intake of total water and water from fluids, and constipation, a low intake of water from foods was associated with an increasing prevalence of constipation. To our knowledge, no previous observational study has investigated the relationship between the intake of water (as a nutrient) and constipation. In contrast, experimental studies have shown that low water intake may be an etiologic factor for constipation, although the potential beneficial effect of extra water intake is unclear (Klauser et al., 1990; Anti et al., 1998; Young et al., 1998; Chung et al., 1999; Arnaud, 2003).
Further, while we are unaware of previous studies examining the association between magnesium intake and constipation, we found that a low intake of magnesium was associated with an increasing prevalence of this condition. Magnesium might form sulfate or citrate salts that would promote fluid retention in the digestive tract and indirectly alter motility, and thereby act as a light laxative (Saez, 1991). The effect of magnesium on constipation warrants further examination.
Several limitations of our study can be identified. First, generally, given that increased intake of dietary fiber and fluid is a widely recommended treatment for constipation (Muller-Lissner et al., 2005), subjects suffering from constipation might be expected to increase their intake of dietary fiber and water (from fluids). Such dietary change is particularly likely in our subjects, who were dietetic students and therefore may have been highly health conscious. In fact, a considerably large percentage of female Japanese dietetic students considered a high intake of dietary fiber and water beneficial for preventing constipation (81 and 42%, respectively) (Ohya and Yoneda, 1995). Thus, the null associations between dietary fiber, total water and water from fluids and constipation observed in the present study might have been due to a possible increase in intake of dietary fiber and water (from fluids) in subjects defined as having constipation. However, because water from foods and magnesium are generally unlikely to be recognized as having an effect on constipation, it is reasonable to consider that our subjects suffering from constipation neither increased nor decreased their intakes of water from foods and magnesium.
Second, all self-reported dietary assessment methods are subject to measurement error and selective underestimation and overestimation of dietary intake (Livingstone and Black, 2003). However, to minimize these possibilities, we used a previously validated DHQ (Sasaki et al., 1998a, 1998b, 2000), although the validity of total water and water from fluid was somewhat insufficient, which might explain the observed null association between these dietary variables and constipation. Additionally, the same tendency of associations between dietary variables and constipation was observed in a repeated analysis of 2717 subjects with a ‘physiologically plausible’ energy intake, namely those possessing a ratio of reported energy intake to estimated basal metabolic rate (standard value of basal metabolic rate for Japanese women aged 18–29 years (99 kJ/kg of body weight/day) multiplied by body weight of each subjects (kg) (Ministry of Health, Labour, and Welfare, 2005)) of 1.2–2.5 (Black et al., 1996) (data not shown). Thus, although the effect of measurement error and selective underestimation or overestimation of dietary intake can never be excluded, data inaccuracy is unlikely to have had a major impact on our findings.
Third, given that our subjects were selected female dietetic students, our results might not be extrapolatable to general populations. Additionally, although we attempted to adjust for a wide range of potential confounding variables, we cannot rule out residual confounding owing to these or poorly measured variables such as physical activity level, which were assessed by a limited number of non-validated questions, as well as other unknown variables.
In conclusion, after adjustment for a variety of potential confounders, low intake of water from foods and of magnesium was independently associated with an increased prevalence of functional constipation among young women whose dietary fiber intake was relatively low. Because the cross-sectional nature of the present study precludes any causal inferences, however, further studies using prospective designs are required to clarify these relationships.
We thank the students for their generous participation in the study. The members of the Freshmen in Dietetic Courses Study II Group (in addition to the authors) are as follows (shown in alphabetical order of the affiliation): S Awata (Beppu University); T Watanabe and A Suzuki (Chiba College of Health Science); T Abe (Doshisha Women's College); H Hayabuchi (Fukuoka Women's University); R Ueda (Futaba Nutrition College); N Takeda and T Matsubara (Hiroshima Bunkyo Women's University); H Ohwada and K Hirayama (Ibaraki Christian University); C Maruyama (Japan Women's University); M Makino (Jin-ai Women's College); S Tanaka and N Nagasawa (Jumonji University); F Tonozuka and S Osada (Junior College of Kagawa Nutrition University); K Uenishi (Kagawa Nutrition University); T Sagara (Kanazawa Gakuin College); Y Enomoto, K Okayama and H Ooe (Kitasato Junior College of Health and Hygienic Sciences); K Nakayama and M Furuya (Kochi Gakuen College); N Yagi and K Soeda (Koshien University); J Ikeda (Kyoto Bunkyo Junior College); I Kitagawa (Kyoto Koka Women's University); K Yokoyama and R Nakayama (Kyoto Women's University); A Miura (Kwassui Women's College); K Baba (Mie Chukyo University Junior College); Y Sugiyama and M Furuki (Minami Kyushu University); T Oyama (Miyagi Gakuin Women's University); Y Naito and M Kato (Mukogawa Women's University); N Hirota (Nagano Prefectural College); T Tsuji and K Washino (Nagoya Bunri University); T Yawata and C Shimamura (Nara Saho College); N Murayama (Niigata University of Health and Welfare); R Watanabe (Niigata Women's College); M Yamasaki (Nishikyusyu University); M Kitamura (Osaka Aoyama College); T Iwamoto (Prefectural University of Hiroshima); I Suzuki and Y Sugishima (Prefectural University of Kumamoto); M Aoki (Sanyo Gakuen College); S Nishi (Seibo Jogakuin Junior College); K Toyama and R Amamoto (Seinan Jo Gakuin University); N Takahashi and R Sasaki (Sendai Shirayuri Women's College); N Kakibuchi (Setouchi Junior College); M Goto (Shokei Gakuin College); M Watanabe and M Yokotsuka (Showa Women's University); M Kimura (Takasaki University of Health and Welfare); M Hara and N Kiya (Tenshi College); J Hirose, T Fukui and K Shibata (The University of Shiga Prefecture); R Nishiyama (Toita Women's College); N Tomita (Tokiwa Junior College); J Oka and T Ide (Tokyo Kasei University); T Uemura and T Furusho (Tokyo University of Agriculture); A Notsu and Y Yokoyama (Tottori College); T Kuwamori (Toyama College); S Shirono (Ube Frontier College); T Goda (University of Shizuoka); K Suizu (Yamaguchi Prefectural University); H Okamoto (Yamanashi Gakuin Junior College).